The Study Of Thin Layer Identification Method Based On Multi-wavelet Spectral Inversion Of Seismic Data

With the increasingly deepening of the current oil and gas exploration,the target of the exploration is gradually transf orming from structural oil a nd gas reservoirs to complex and subtle lith ologic reservoirs.Because of the complexity of the continen tal basin sedimentary environm ent and the extens ive development of the thin layer,the effective identification of the thin layer sand body is the key to the oil and gas exploration.However,it's difficult to identify the thin layer sand body as a result of the seismic resolution limitation.For this reas on,carrying out the study of the m ethod of multiav-welet spectral invers ion of seis mic data under th e goal of improving the resolution of the seismic profile in order to achieve the effective recognition of the thin layer.Firstly,the spectral characteristics of the thin layer reflection coefficient are studied in the paper by theoretical m odels,initially recognizing the spectra l characteristics of the thin layer reflection coef ficient,which lays a foundation for the following study of the spectral inversion theory.And then th e expression of the thin layer reflection coefficient is derived in the frequency do main and the inv ersion objective function is obtained based on this expression.In addition,the theory of signal parity-decomposition is introduced into the spectral inversi on objective function.A ccording to the good ability to recognize the thin la yer of the even com ponent in the reflection coef ficient,appropriately increasing the weight of ev en component to im prove the recognition ability of the thin layer.Furthermore,it's the multi-wavelet decomposition of seismic data.In view of the shortcoming of computational efficiency of the traditional matching pursuit algorithm,the strutured matching pursuit algorithm is proposed to decompose the seismic signal into multi wavelets.T his algorithm can reduce the computational c omplexity and improve the efficiency of matching decomposition by using the method of packet-mode atom library and tree-structured search.The va lidity and r eliability of multi-wavelet decomposition based on the structured m atching pursuit are analy zed on the basis of model testing,achieving the goal of deco mposing the seism ic signal into the combination of di fferent wavelets m atching with the s ignal feature.Moreover,the accuracy of this multi-wavelet decomposition based on the structured matching pursuit is verified by recons tructing the wavelets after decomposition to res tore the s eismic signal.The spectrum of the complex reflection co efficient sequence is relatively complex and the efficiency of the process in spectral inversion is low.In view of these problems,combined with the structured m atching pursuit decomposition,the thin layer reflection coefficient spectral inv ersion on multi wave lets after decom position is proposed to improve the inversion efficiency.In the application of theoretical models and actual data,this method of multi-wavelet spectral inversion is verified to improve the resolution of the seismic profile effectively and enhance the reflection information of the thin lay er under ground,which can provide effective guidance for oil and gas exploration.